Inflammatory diseases of the parathyroid gland

Talat N, Diaz‐Cano S & Schulte K‐M (2011) Histopathology59, 897–908 Inflammatory diseases of the parathyroid gland


Introduction
Inflammatory disorders of the parathyroid gland are rare as compared with those of other endocrine organs. [1][2][3][4] These entities are poorly defined, and descriptive terms such as lymphocytic infiltrate, parathyroiditis and lymphocytic parathyroiditis are used, resulting in no agreed scheme of classification.
Inflammation in other endocrine organs such as the thyroid has a long list of notable clinical associations. Thyroiditis, for example, is known to be associated with hypothyroidism, goitres, thyrotoxicosis, and hashitoxicosis, 5,6 but no such clinical associations have been defined for parathyroid inflammatory disorders. Thyroiditis is associated with severe disturbances in thyroid function, with inflammatory effects ranging from thyrotoxicosis resulting from hormone releasing thyroiditis to chronic subclinical and clinical hypothyroidism, when the gland is transformed into a scar. With regard to parathyroid inflammation, very little is known about its effects on parathyroid function. A number of studies have demonstrated autoimmune hypoparathyroidism; however, none has shown histological evidence of inflammatory processes within the parathyroids as is seen in parathyroiditis. [7][8][9] Because of this terminological variation and the paucity of recent studies, it is not generally clear how to classify inflammatory parathyroid entities. It is also not clear whether such descriptive nosological entities relate to specific clinical constellations. This article is aimed at clarifying the pathological presentations and their clinical associations by using findings from an index patient with hyperplastic parathyroiditis; the molecular markers of this lesion are compared with a set of normal parathyroid glands, hyperplasias, and adenomas. We have therefore attempted to improve our understanding of these rare conditions by a review of the literature. It also makes minor contributions to our understanding of functional aspects of inflammatory parathyroid disease.

Materials and methods
A systematic search was conducted of the published literature on inflammatory diseases of the parathyroid gland. The databases searched included PubMed (http://www.ncbi.nlm.nih.gov/pubmed/), Thompson ISI (http://apps.isiknowledge.com/) and Google Scholar; the search terms used were 'parathyroiditis', 'inflammation of parathyroid gland', 'lymphocytic infiltrate', 'tuberculosis', 'sarcoidosis', and 'granulomatous inflammation'. A hand search of article bibliographies was also performed with the ISI Thompson Web of Knowledge Citation report, enabling further articles to be retrieved. From the initial electronic search of 185 articles for possible inclusion, 27 articles were retained following title and abstract reviews; these articles reported on a total of 96 patients with inflammatory diseases of the parathyroid gland, including parathyroiditis (n = 15), lymphocytic infiltrate (n = 69), sarcoidosis of the parathyroid gland (n = 6), tuberculosis (n = 14), and granulomatous disease (n = 2).  The inclusion criterion was unambiguous evidence of inflammatory processes of the parathyroid gland in the form of a histological description.
Histopathological descriptions were evaluated from the author's summary diagnosis and by analysing the information provided by the authors regarding the presence of lymphocytic infiltrate, parathyroiditis, and granulomatous inflammation. The terminology for parathyoiditis was inconsistent. After review of all histological descriptions, we found two fundamentally different patterns, and grouped patients according to these descriptions: 1. The non-specific pattern is marked by diffuse lymphocytic infiltrates in the direct vicinity of venules without evidence of lymphocyte maturation or immune-mediated tissue damage, such as fibrosis or epithelial degeneration. 2. The pattern of lymphocytic parathyroiditis is characterized by interstitial lymphocytes away from the vessels with terminal differentiation (plasma cells) and ⁄ or formation of germinal centres. Data were extracted by one researcher and checked by another with standardized extraction tables developed a priori. Data were pooled as individual cases in a single contingent spss table.
patient This 42-year-old male Caucasian patient suffered from end-stage renal failure resulting from hyperplastic kidney that he had suffered from since 1995, when he was put on continuous ambulatory peritoneal dialysis. He developed severe secondary hyperparathyroidism with fatigue and bone pain, failed on calcimimetics, and underwent a total cervical parathyroidectomy with central lymph node clearance and thymectomy in 2007. Histopathology demonstrated four parathyroid glands with nodular hyperplasia, weighing 4.96 g, 1.64 g, 1.23 g, and 0.71 g. Two lymph nodes showed only reactive changes; the thymus and mediastinal fat resection weighed 125.5 g, and were histologically normal. Postoperatively, the parathyroid hormone (PTH) level dropped to 97 ng ⁄ l (normal 13-75 ng ⁄ l) in the presence of a low normal corrected calcium level of 2.29 mm and a normal phosphate level ( Figure 1). The presence of PTH at this point was explained by the 60-mg parathyroid autotransplant into the forearm extensor muscles. An increase in the PTH level to >10-fold upper normal levels, with recurrent fatigue and bone aches, led to excision of the autotransplant from the left forearm after negative imaging findings in the neck and chest. This revealed 1.99 g of nodular parathyroid hyperplasia. The postoperative PTH levels dropped to 52 ng ⁄ l, in the presence of a normal corrected calcium level supported by high-dose active vitamin D treatment. Over the subsequent 2 years, the patient showed extreme fluctuations of PTH levels between the upper normal limit of 70 ng ⁄ l and 2000 ng ⁄ l with concordant fluctuations in symptomatology. Complex reimaging with ultrasound and magnetic resonance imaging of the neck, contrast computed tomography (CT) of the neck and chest and fusion imaging including Technetium Tc 99m sestamibi single-photon emission computed tomography (MIBI-SPECT) sug-gested a 1-cm lesion riding the aortic arch. In July 2010, initial thoracoscopic exploration showed this to be a probable tuberculous lymph node, and surgery was therefore changed to thoracotomy with exploration of the aortopulmonary window. The surgery was followed by normalization of the PTH and calcium levels. The specimen measured 35 · 19 · 21 mm and weighed 6 g. It showed a nodular growth pattern composed of oncocytic and clear cell islands. The stroma was expanded by a prominent diffuse and patchy mixed inflammatory infiltrate, comprising mainly plasma cells and lymphocytes. There was interstitial haemorrhage, congestion, and large numbers of haemosiderin-laden macrophages. There was no vascular invasion, confluent necrosis, or undue mitotic activity. The features were consistent with hyperplastic parathyroiditis (Figures 2 and 3).

targeted gene expression analysis
Total RNA was isolated from control parathyroid glands (n = 5, from thyroidectomies with no parathyroid pathology), hyperplasias (n = 10), and adenomas (n = 10), by the use of commercial kits (RNeasy kit; Qiagen, Crawley, UK), following the manufacturer's instructions. After extraction, the RNA was treated with DNaseI (DNA-free; Ambion, Austin, TX, USA) at 37°C for 20 min in a 50-ll reaction containing total RNA in water, 5 ll of DNA-free buffer, 2 ll of Superase   RNAse inhibitor (Ambion), and 1 ll of DNaseI. To clean up the RNA, the mixture was loaded onto an RNeasy column and total RNA was eluted with water in a volume of 40 ll. The quantity and quality of the RNA samples were determined with the RNA 6000 LabChip kit (Agilent Technologies, Wilmington, DE, USA). Real-time quantitative polymerase chain reaction (PCR) Quantitative real-time PCR is based on the detection of a fluorescent signal that increases linearly with accumulating amplification product during the PCR reaction. During the extension phase of PCR, the probe is cleaved by the endogenous 5¢-nuclease activity of AmpliTaq Gold polymerase, which cleaves the reporter chromophor from the TaqMan probe, leading to an increase in the intensity of reporter fluorescence. The increase in fluorescence (DRn) signal is continuously measured and plotted versus PCR cycle number, reflecting the amount of PCR amplification product. The abi prism 7700 detection system software (Applied Biosystems, Foster City, CA, USA) calculates DRn by subtracting the fluorescence signal of the baseline emission during cycles 3-6 from the fluorescence signal of the product at any given time. A threshold was set at the early log phase of product accumulation. The threshold cycle number value (C T ) is the cycle number at which each sample's amplification plot reaches this threshold. We carried out real-time PCR with TaqMan unlabeled PCR primers and 6-carboxyfluorescein dyelabelled TaqMan minor groove binder (MGB; 6-carboxytetramethylrhodamine) primer probes on an abi prism 7700 Sequence Detection System (Applied Biosystems). The MGB probes have an MGB at the 3¢-end of the probe, which increases the T m of probes, allowing the use of shorter probes. PCR primers and fluorogenic probes (TaqMan) were designed by using Primer Express software (PE Biosystems). All primer sequences are given in Table 1. Wherever possible, primers and probes were designed to span introns in the genomic DNA, thereby minimizing the potential for confounding of the signal by contaminating genomic DNA. TaqMan primers and MGB probes for target genes (Table 1) were obtained from Assays on Demand (Applied Biosystems). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) TaqMan primers and MGB probes were obtained as pre-developed assay reagents from Applied Biosystems. PCR reactions were performed in 25-ll volumes consisting of 1· PCR buffer A (PE Biosystems), 3.5 mm MgCl 2 , 300 lm each dATP, dCTP, dGTP, and dTTP, 1.25 units of AmpliTaq Gold (PE Biosystems), and 2.5 ll of the appropriate RT reaction product. The primers were present at 300 nm, and the TaqMan probe was present at 200 nm. The GAPDH amplification primers were each present at 100 nm, and the TaqMan probe was present at 100 nm. Amplification and detection were carried out by using an abi 7700 detection system as follows: one cycle at 95°C for 12 min, 40 cycles at 95°C for 15 s, and 60°C for 1 min. All cases were conducted with duplicate samples and C T units obtained as the average of the results from the replicates. C T is defined as the threshold cycle for PCR amplification that is detected by the abi prism DNA sequence detector at a threshold value set at 0.45 for each run. The calculated C T values were analysed in Microsoft Excel.

Calculation of relative expression
Relative mRNA expression of target genes was calculated with the comparative C T method. The amount of target gene was normalized to the endogenous GAPDH control gene to control for quantity of RNA input. The difference in C T values was calculated for each mRNA by taking the mean C T of duplicate reactions and subtracting the mean C T of duplicate reactions for reference RNA measured on an aliquot from the same RT reaction: DC T = C T (target gene) -C T (GAPDH). A logarithmic transformation was used to graphically tabulate the data.

Results
There were two large autopsy studies specifically addressing the issue of parathyroiditis, one comprising 225 patients 10 and one containing 589 patients, 36 in whom 1637 parathyroid glands were dissected. Both studies were based on unselected autopsy cohorts, with male patients accounting for 51.1% and 57.0%, respectively. Seemann's study revealed two patients There was a single case of disseminated tuberculosis. This patient also showed caseating granulomas within the parathyroid gland, along with multiple perivascular lymphocytic infiltrates. This article's conclusions highlighted only one patient with true lymphocytic parathyroiditis in a patient with myocardial infarction, and proposed an autoimmune process. The Thiele study showed 10 patients with secondary hyperparathyroidism, but no cases of primary hypoparathyroidism ⁄ hyperparathyroidism.
It focused mostly on the problem of transition between diffuse and nodular hyperplasia and adenoma. The article described four patients with purulent destruction of parathyroid parenchyma in the setting of septicaemia or continuity from infected tracheostomies. It also described diffuse lymphocytic infiltrates in an unknown number of patients, but these findings were dismissed. The article concluded that no single case of true lymphocytic parathyroiditis was observed. The results of the autopsies studies are summarized in Table 2. Table 3 gives clinical details of the 15 cases of parathyroiditis identified from the literature. All patients included had clear histological evidence of the disease, as described above. We reclassified three cases, described by the authors as lymphocyic infiltrate of the parathyroid gland, as histologically consistent with parathyroiditis. These patients were therefore reclassified as having parathyroiditis. Apart from the autopsy studies, all other patients showed parathyroid hyperfunction.
In surgical pathology material, only five case reports mentioned the presence of lymphocytic infiltrate of the parathyroid gland in patients with primary hyperparathyroidism. All patients (5, 100%) were found to have a parathyroid adenoma (Table 4). Associated diseases were found in two of four (50%; one case had no data available). Associated diseases were far more common in the single autopsy cases (5 ⁄ 7, 71%) ( Table 4). Table  6 shows that associated diseases were much more common in patients with simple lymphocytic infiltrate of the parathyroid gland (73.9%) than in patients with parathyroiditis (31.2%) (v 2 11.1, P = 0.001). Granulomatous disease of the parathyroid gland is surprisingly rare. Only 11 cases have been identified, and the clinical details of these are outlined in Table 5.
The relative gene expression for each marker is shown in Figure 4. All markers for our case of hyperplastic parathyroiditis were in the range of hyperplastic-adenomatous lesions of the parathyroid gland, with the exception of CD68, which was significantly upregulated, reflecting the presence of histiocytic infiltrate [also supported by mannose 6-phosphate receptor (M6PR) upregulation]. Hypoxia-inducible factor 1A (HIF1A) and malate dehydrogenase 2 (MDH2) upregulation suggested an ischaemic background and a reactive mitochondrial process. The upregulated carbonic anhydrase 4 (CA4) expression is consistent with hyperparathyroidism.

Discussion
The term parathyroiditis has been used with some inconsistency. Some authors use it to describe any lymphocytic infiltrate in the parathyroid gland, whereas others employ it as a specific term for evidence of an immune process residing in the parathyroid gland. 36 Other authors use the term lymphocytic infiltrate of the parathyroid gland to characterize what other authors would have described as parathyroiditis. [23][24][25] Because of this terminological variation and the paucity of recent studies, it is not generally clear how to classify inflammatory parathyroid entities.
We start from the general observation that any lymphocytic infiltrate defines a primary inflammatory condition of the involved organ. Genuine organ-related inflammation requires evidence of an organized immune process, e.g. in germinal centres, and ⁄ or evidence of immune-mediated cellular or interstitial damage. There are two fundamentally different patterns of distribution of immune cells in the parathyroid gland. The non-specific pattern is marked by perivenular lymphocytic infiltrates without evidence of lymphocyte maturation or immune-mediated tissue damage, such as fibrosis or epithelial degeneration. The pattern of lymphocytic parathyroiditis is defined by interstitial chronic inflammation with mature plasma cells and ⁄ or germinal centers, along with evidence of epithelial reaction (either degenerative or proliferative). Our reported case clearly documents these defining elements (Figures 2 and 3), along with molecular markers supporting a histiocytic component (CD68 upregulation) in an ischaemic background (HIF1A upregulation) with mitochondrial reaction (MDH2 upregulation) and hyperparathyroidism (CA4 upregulation). In light of this definition, we have reviewed all published cases. Lymphocytic infiltrates are not uncommon, and have been identified in 6% of nearly 1000 autopsy cases in three large series of unselected patients (Tables 2 and 4). They do not seem to relate to functional hyperparathyroidism or hypoparathyroidism, and the autopsies have not identified organ damage related to parathyroid dysfunction. In contrast to this frequent presentation, genuine parathyroiditis is rare (only 15 cases, four of which were identified among 814 autopsy cases, accounting for a total of no more than 0.5%) (Tables 2-4). If the three cases reported in autopsy case reports are excluded from statistical analysis, the frequency drops to 0.1%. We can safely conclude that genuine parathyroiditis is a very rare condition, on the basis of both autopsy series and the low number of surgical specimens. These surgical specimens were obtained with the aim of curing primary and secondary hyperparathyroidism. We think that the association with parathyroid hyperfunction is spurious and is related to the sampling mechanisms, as parathyroid removal is almost always performed in the setting of hyperfunction.
We conclude that these two patterns are not only morphologically distinct but also pathogenetically different. Lymphocytic infiltrates of the parathyroid gland appear to be statistically associated with systemic disease; this observation is significantly less common in parathyroiditis. We propose that the basis of this finding is directly related to the origin of the lymphocytic infiltrate. It is clearly associated with two conditions that would generally impact on venule integrity of the parathyroid gland, the neck, or beyond. Septicaemia and septic shock are characterized by a loss of integrity of the small vessels caused by bacterial toxins and excessive production of immune mediators by the human body. Depending on the degree of damage, the capillary leak progresses from simple oedema to massive loss of intravascular protein, lymphocytic extravasation, and finally microhaemorrhage into multiple organs. 37,38 The other condition associated with perivenular diffuse lymphocytic infiltrates is myocardial infarction. Left heart failure leads to right heart failure, venous congestion, increased hydrostatic pressure, and secondary disruption of venule integrity, especially when reduced oxygen delivery enhances tissue vulnerability ( Table 6).
The above concept provides significant help in understanding the more frequent non-specific pattern, but it does not shed any light on the origin of the primary parathyroid immune process. The nature of the current data does not invite any related speculations. Perhaps the most remarkable observation of this study is the rarity of specific inflammatory processes in the parathyroid gland. This in striking contrast to the neighbouring endocrine organ, i.e. the thyroid, where autoimmune or perhaps infection-related thyroiditis is observed with a lifetime incidence of 10%. 2,5 For some reason, the parathyroid gland is either poorly antigenic or protected against the processes that initiate so-called autoimmunity. The latter observation could become relevant in the perspective of transplantation of small cell volumes.     We also identified 14 cases of granulomatous inflammation of the parathyroid gland, including six cases of sarcoidosis, four cases of tuberculosis, and two nonspecific granulomatous processes (Table 5). Again, the rarity of this association warrants attention when seen on the background of thousands or millions of described cases of granulomatous processes in other tissues. 1,39,40 We agree with Furuto-Kato 15 that parathyroiditis may be associated with significant functional aberrations ( Figure 1); an explanation of this observation is offered by molecular studies showing the impact of inflammatory substances on functional regulation of the parathyroid gland. 41,42 Such immune mediators could plausibly be liberated and act as paracrine regulators in the vicinity, for example, of germinal centres of the parathyroid chief cell mass.

Conclusion
Genuine inflammatory disorders of the parathyroid gland are rare, and may result in functional dysregulation. The term parathyroiditis should be reserved for lesions with evidence of a primary parathyroid immune process. The common lymphocytic infiltrate is a non-organ-specific reflection of a capillary permeability change and, at best, indicates clinical disease. The diagnosis of lymphocytic infiltrate as such should not lead to any other examinations or diagnostic procedures. The rarity of granulomatous disease in the parathyroid gland points to specific immunological properties of the parathyroid gland.  Inflammatory diseases of the parathyroid gland 907